A connection may be made between the number of cells and the number of inhabitants of the police district, at least in towns where crime is, for the most part, local (therefore excluding Chessy and Saint-Ouen for the reasons already given). This does not constitute a rigorous approach, since some crime originates outside of the police district. Above all, as mentioned above with regard to the level of persons implicated, the differences bring other factors into play, in particular the scale of police activity. However, the cell/inhabitants ratio at least provides a sign of the manner in which the necessary facilities were thought out within the managing bodies. As might be expected, there are considerable differences: varying between one cell to 4,400 inhabitants (at Troyes) and one for 34,200 inhabitants (at Amiens), that is to say a difference of 1 to 8. The geography of distribution does not necessarily coincide with high crime figures. The best-equipped towns, apart from Troyes, are Montereau-Fault-Yonne, Pointe-à-Pitre and SaintQuentin; the most poorly-equipped, apart from Amiens, are Angers, Niort and Saint-Denis. The logic of real estate opportunities and reconstructions presumably prevailed over that of adaptation of means to crime levels. This is understandable, but also to be regretted. Naturally, a connection needs to be made between the number of cells and the daily number of placements in police custody, in order to shed light upon phenomena of cell overcrowding if necessary. Admittedly, a connection of this kind is somewhat tenuous, since placements in police custody last for under twenty-four hours and the same cell may therefore be occupied two or three times in succession in the space of twenty-four hours. However, the above data establishes that the average length of police custody is in the order of 17 hours and that the factor of shorter placements needs to be weighed against the need to separate the sexes and different age groups as well as the requirements of investigations. Furthermore, as already mentioned, additional fluxes may be generated by persons originating from other police stations. It is therefore not unreasonable to establish a parallel between the respective magnitudes of these two sets of figures. The towns in which the average daily number of placements in police custody is greater than the number of available cells (not including sobering-up rooms) are Amiens, Angers, Metz and Saint-Denis. In all of these towns, the police station contains collective cells, which it may therefore be assumed may often contain considerable numbers of people. At Vénissieux and Villeneuve-Saint-Georges, the numbers (placements in police custody and cells) are very close but there are no collective cells: considerable congestion in small cells may occur, at least at certain periods. Finally, the data is optimal. Damage and the slowness of repairs (electricity, plumbing, door and window frames), which is everywhere observed, means that certain cells are permanently taken out of service because they represent a threat to personal safety (e.g. exposed electrical wires). A certain number of existing cells are therefore unavailable, which may heighten the risks of congestion. b/ The second criterion concerns the “comfort” of cells. Two elements are here taken into consideration. On the one hand, the existence of absence of toilets (in the exclusive form of toilet bowls, without seats; one should not set one’s hopes too high), which are present in police stations built or renovated over the last twenty years or so. In principle, these toilets are placed behind a low wall intended to protect privacy. They are usually accompanied with a water supply built into the wall directly above the toilet. Such facilities still account for a minority of cases: less than a quarter of the police stations in the sample. In the others, persons in police custody are obliged to call if obliged by necessity, with more or less happy results. In the first place their call has to be heard and secondly an officer needs to be available. Dirtiness of cells arises from provocation; but also from urgent and unsuccessful requests to come out of the cell, as the testimonies show. 37